Association between anthropometric indices and chronic kidney disease: Insights from NHANES 2009-2018

Abstract

Introduction: The strong association between obesity and chronic kidney disease (CKD) has been empirically validated, yet traditional measures like the Body Mass Index (BMI) fail to accurately assess the extent of obesity due to CKD's characteristics, such as reduced muscle mass and increased visceral fat. This study investigates the association between CKD and several anthropometric indices, including A Body Shape Index (ABSI), Body Roundness Index (BRI), Waist-to-Height Ratio (WHtR), and the Conicity Index (C-index), to determine their predictive capabilities.

Methods: Based on the datasets from the National Health and Nutrition Examination Survey (NHANES) 2009-2018, weighted multivariable regression analyses were carried out to examine the independent relationship between two anthropometric indices and CKD. Also, subgroup analyses, restricted cubic spline regression (RCS), and receiver operating characteristic curve analysis were conducted for further data analyses.

Results: A total of 24,162 participants were enrolled in this study. After adjusting for confounding factors, ABSI, BRI, WHtR, and the C-index were significantly associated with an increased risk of CKD, while BMI was not. Height showed a protective effect against CKD. ABSI and the C-index demonstrated the highest areas under the curve (AUCs), indicating superior predictive capabilities compared to traditional measures like BMI and waist circumference (WC). Subgroup analyses revealed significant interactions between the anthropometric indices and factors such as age, disease status, dietary intake, and physical activity levels.

Conclusions: This study highlights the significant associations between various anthropometric indices (including ABSI, BRI, WHtR, and C-index) and the risk of CKD. ABSI and the C-index demonstrated the strongest predictive capabilities for CKD, with the highest AUC values.

Fig 1. A flowchart showing the selection of study participants.

Fig 2. Pairwise Pearson correlations among anthropometric indices.

Anthropometric indices were sex-specific z-scores transformed. BMI, Body Mass Index; WC, waist circumference; ABSI, A Body Shape Index; BRI, Body Roundness Index; WHtR, Waist-to-Height Ratio; C-index, Conicity Index.

Fig 3. The dose–response association between anthropometric and CKD.

The dose-response relationship was evaluated using restricted cubic spline regression for WC(a), BMI (b), ABSI (c), and C-index (d) and CKD, with covariates adjusted as in model 3. The odds ratio is represented by the red line and the 95% confidence interval is shown in pink.

Fig 4. Association between the sex-specific z-score of ABSI and CKD in subgroup and tests for interaction, with covariates adjusted as in model 3.

* p < 0.05. The number of CKD cases per category for categorical variables is as follows: for age, there are 1448 cases in the <60 group and 2686 cases in the ≥60 group; for race, there are 937 cases among Black individuals and 3197 cases among others; for hypertension, there are 1507 cases with no hypertension and 2627 cases with hypertension; for diabetes mellitus, there are 2427 cases without diabetes and 1707 cases with diabetes; for cardiovascular disease, there are 3068 cases without cardiovascular disease and 1066 cases with cardiovascular disease. For continuous variables, the number of CKD cases per group is as follows: for diet energy, there are 1755 cases in the <1622 kcal/day group, 1346 cases in the 1622–2255 kcal/day group, and 1033 cases in the ≥2255 kcal/day group; for MET, there are 2194 cases in the <600 min/week group, 1245 cases in the 600–4000 min/week group, and 695 cases in the ≥4000 min/week group.

Fig 5. Association between the sex-specific z-score of C-index and CKD in subgroup and tests for interaction, with covariates adjusted as in model 3.

* p <0.05.

Fig 6. ROC curves for the prediction of CKD by anthropometric indices, weighted.